Skip to content
FeltnerAI

Set Up a Server

This page covers standing up a FeltnerAI server from a release build. If you just want to try it locally, start with the Quick Start.

Build a release server

Section titled “Build a release server”

The web app must be built first so it is embedded into the server binary.

  1. Install frontend dependencies:

    Terminal window
    bun install --cwd frontend --frozen-lockfile

  2. Build the web app:

    Terminal window
    bun run --cwd frontend build

  3. Build the release server:

    Terminal window
    cargo build --release -p feltnerai-server

The resulting target/release/feltnerai-server binary serves both /api/v1 and the web app. On Windows it also runs a system-tray menu with Open in browser and Exit actions.

First-run setup

Section titled “First-run setup”

  1. Start the server. It prints a temporary setup token to the console.

  2. Open the server in a browser (default http://127.0.0.1:8080).

  3. Complete the guided setup:

Create user accounts

Section titled “Create user accounts”

FeltnerAI has no open registration. Administrators create accounts from Admin → Users. Each new account is forced to change its password on first sign-in.

From the admin UI you can manage:

  • Users — create, disable, and reset accounts
  • Providers — add OpenAI-compatible providers and test connections
  • Models — configure which discovered models are available
  • Server — settings, start-at-login (Windows), and backup/restore
  • Branding — logo, favicon, and custom CSS

Start at login (Windows)

Section titled “Start at login (Windows)”

On Windows, administrators can toggle Open FeltnerAI Server when I sign in to Windows from Admin → Server. This registers the current executable under the current user’s standard Windows Run key.

Next steps

Section titled “Next steps”